I'm searching for a pseudo filesystem to use instead of raid/lvm on a bunch of disks.
The problem is, that one is either waisting some space (for ex. raid5) or is reducing fault tolerance (loose one disk in JBOD setup, loose all data).
In situations where contents aren't critical, elegant solution seems to be overlay filesystem:
any number of disks (or partitions, each one already has fs (xfs, ext2 etc.)) would be "gathered" in one virtual, so that directory structure is identical on all of them, but files are distributed.
Anyone heard of such thing implemented (it seems rather trivial to code) or has comments?
The problem is, that one is either waisting some space (for ex. raid5) or is reducing fault tolerance (loose one disk in JBOD setup, loose all data).
In situations where contents aren't critical, elegant solution seems to be overlay filesystem:
any number of disks (or partitions, each one already has fs (xfs, ext2 etc.)) would be "gathered" in one virtual, so that directory structure is identical on all of them, but files are distributed.
Anyone heard of such thing implemented (it seems rather trivial to code) or has comments?
Facebook
Twitter